EUVIMEDEUVIMED entdecken
Crossref

Subsampled trust-region methods with deterministic worst-case complexity guarantees

Max L N Gonçalves, Geovani N Grapiglia

IMA Journal of Numerical Analysis · 10.11.2026

DOI: 10.1093/imanum/drag111 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IMA Journal of Numerical Analysis, online erschienen am 10.11.2026. Verfasst von Max L N Gonçalves und 1 weiteren Autor:innen. Die Arbeit nennt 26 Literaturangaben. Gefördert von: FAPEG, Fonds d’appui l’internationalisation.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

In this paper, we develop and analyse subsampled trust-region methods for solving finite-sum optimization problems. These methods employ subsampling strategies to approximate the gradient and Hessian of the objective function, significantly reducing the overall computational cost. We propose a novel adaptive procedure for deterministically adjusting the sample size used for gradient (or gradient and Hessian) approximations. Furthermore, we establish worst-case iteration complexity bounds for obtaining approximate stationary points. More specifically, for a given $\varepsilon _{g}, \varepsilon _{H}\in (0,1)$, it is shown that an $\varepsilon _{g}$-approximate first-order stationary point is reached in at most $\mathscr{O}({\varepsilon _{g}}^{-2} )$ iterations, whereas an $(\varepsilon _{g},\varepsilon _{H})$-approximate second-order stationary point is reached in at most $\mathscr{O}(\max \{\varepsilon _{g}^{-2}\varepsilon _{H}^{-1},\varepsilon _{H}^{-3}\})$ iterations. Finally, numerical experiments illustrate the effectiveness of our new subsampling technique.

Förderung

  • FAPEG
  • Fonds d’appui l’internationalisation

Zitieren

APA

Gonçalves, M. L. N., & Grapiglia, G. N. (2026). Subsampled trust-region methods with deterministic worst-case complexity guarantees. IMA Journal of Numerical Analysis. https://doi.org/10.1093/imanum/drag111

Vancouver

Gonçalves MLN, Grapiglia GN. Subsampled trust-region methods with deterministic worst-case complexity guarantees. IMA Journal of Numerical Analysis. 2026. doi:10.1093/imanum/drag111

BibTeX
@article{Gonalves2026,
  author = {Gonçalves, Max L N and Grapiglia, Geovani N},
  title = {Subsampled trust-region methods with deterministic worst-case complexity guarantees},
  journal = {IMA Journal of Numerical Analysis},
  year = {2026},
  doi = {10.1093/imanum/drag111}
}
Kostenlos starten →